INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.07

2:0.09

3:0.09

4:0.08

5:0.09

6:0.08

7:0.07

8:0.07

9:0.07

10:0.08

11:0.08

Negative Logits

oots

-1.62

iasis

-1.43

ippi

-1.33

unes

-1.32

atinum

-1.31

ccess

-1.29

 Verb

-1.26

undle

-1.26

 Archives

-1.25

estyles

-1.25

POSITIVE LOGITS

Kry

1.59

pow

1.38

 predatory

1.29

ertodd

1.27

 Bauer

1.22

 decentralized

1.19

 orderly

1.19

 differentiation

1.15

 exponent

1.15

shut

1.14

Activations Density 0.000%

No Known Activations

This feature has no known activations.